Introduction

Root-knot nematodes are one of the most important pests, which causes considerable economic losses worldwide (Oka et al. 2000; Westphal 2011). For decades, several chemical pesticides are being used to control them, however, continued and excess use had led to resistance in the nematodes. Hence, use of potential biocontrol agents to combat parasitic nematodes population is one of most reliable and safe alternative. Nematophagous fungi are the predators of nematodes which can easily trap and devour them completely. Based on the infection processes, these fungi can be classified into three different groups i.e., nematode-trapping, endo-parasitic and eggs or cysts-parasitic. Nematode-trapping fungi forms trapping devices of diverse morphology to capture the nematodes which can be either adhesive network, hyphae, branches, knobs and non-constricting or constricting rings (Ahren and Tunlid 2003; Schmidt et al. 2007; Yang et al. 2007d). Among these, net forming fungi are of special interest and have been extensively studied (Braga et al. 2014; Niu and Zhang 2011; Tunlid and Ahren 2011). The unique nature of these fungi draws attention of researchers to study and use them as biocontrol agents against plant and animal parasitic nematodes.

Arthrobotrys conoides and Duddingtonia flagrans both are nematode-trapping fungi, belongs to order Orbiliales and family Orbiliaceae of Ascomycota and they trap the prey nematodes using their adhesive hyphal network (Drechsler 1937, 1941; Duddington 1949). A. conoides kills both plant as well as animal parasitic nematodes (Al-Hazmi et al. 1982; Araujo 1998; Kalele et al. 2010) and its nutrient requirements (Coscarelli and Pramer 1962; Grant 1962) and morphology (Falbo et al. 2013) is well recognized. On the other hand, D. flagrans controls gastrointestinal parasites of animals (Gronvold et al. 1993; Hertzberg et al. 2001), however, there are no evidences for its predacious activity against plant parasitic nematodes. In one of our previous studies, we noticed that A. conoides (isolate RPAN-12) efficiently traps and kills root-knot nematodes than D. flagrans (isolate RPAN-10) (Pandit et al. 2014b) and this created curiosity to understand the underlying molecular mechanism for this process. Till now only serine protease (AcI) of A. conoides is known to be involved in the infection of nematodes (Yang et al. 2007a). Additionally, recent research on A. conoides also explained the morphological plasticity using the electron microscopy (Falbo et al. 2013).

It is very essential to understand the molecular mechanisms of organisms in order use them as an effective biocontrol agents (Davies 2005; Morton et al. 2004). Recent advances in sequencing technology, especially next generation sequencing (NGS) has opened new horizons in this area. Genome and transcriptome analysis of numerous fungi are available, but there are limited reports for nematode-trapping fungi (Ahren et al. 2005; Andersson et al. 2014; Meerupati et al. 2013; Yang et al. 2011b). Moreover, metabolites obtained from nematophagous fungi have also been identified, which expressed nematicidal activity (Degenkolb and Vilcinskas 2016a, b). However, similar information for A. conoides is missing. RNA-Seq is considered as robust methodology to study the differential gene expression of an organism under different set of conditions. In the present study, we analyzed the differentially expressed genes in A. conoides and D. flagrans after inducing them with Meloidogyne sp. nematode mixture, with the major focus on A. conoides. Understanding the genes associated with the virulence are very helpful for further improvement of the strains for their predacious activities.

Materials and methods

Microorganism and culture conditions

The fungal isolates, Arthrobotrys conoides RPAN-12 (GenBank accession No. JX979095) and Duddingtonia flagrans RPAN-10 (GenBank accession No. JX979094) used in the present study were originally isolated from the soil nearby Anand Gujarat (India), identified and characterized (Pandit et al. 2014b). As reference genome of A. conoides, is not available, transcriptomes of both fungi were sequenced under two different conditions, i.e., in the presence and absence of nematode homogenous mixture to prepare in-house reference file for RNAseq analysis. For this, 8 mm diameter agar blocks containing mycelia of fungi (previously grown at 28 ± 1 °C for 7 days on Corn Meal Agar) were inoculated on the Petri plates (ten for each fungus) of half strength (0.85%) Corn Meal Agar (CMA) (Hi-media, India) to provide nutrient deprived conditions and incubated at 28 ± 1 °C for 4 days. Meloidogyne sp. of root-knot nematodes were collected from the infected tomato plants (Department of Nematology, Anand Agricultural University, Anand, Gujarat) and crude homogenised mixture was prepared in phosphate buffer saline (PBS), pH 7.0 by grinding with sterilized mortar pestle. On 5th day, 1 ml of this mixture was added to five Petri plates of each fungus (induced) and remaining five Petri plates served as control (uninduced), which were flooded with 1 ml of PBS. All the Petri plates were further incubated at 28 ± 1 °C in dark condition for 24 h. Before RNA extraction, plates were observed with the help of light microscope (Leica DM2500) using the 10× objective to evaluate the changes between induced and un-induced fungi.

Extraction of total RNA and purification of mRNA

After 24 h, mycelia were collected from each treatment, i.e. five plates of each fungi and approximately 1/3 of mycelia were preserved in RNAlater (Qiagen, UK) and stored at −80 °C which were later on used for qRT-PCR. Form rest of the mycelia total RNA was extracted using TRIzol (Invitrogen, UK) and RNeasy Mini kit (Qiagen, UK). Total RNA was treated with RNase free DNase (Qiagen, Germany) to remove any DNA traces. Before processing for mRNA extraction, quality (integrity) of RNA was assessed using Agilent 2100 Bioanalyzer (Agilent Technologies) with the help of RNA 6000 nano kit and quantity was measured using Qubit 2.0 fluorometer (Invitrogen, USA). mRNA from the total RNA was purified using the Dyna beads Oligo (dT)20 probes (mRNA isolation kit, Roche).

Library preparation and RNA sequencing

The purified mRNA (~500 ng) was further fragmented using RNase III and size distribution of fragmented mRNA was accessed on the Agilent 2100 Bioanalyzer using DNA High Sensitivity DNA Analysis kit. After fragmentation, purified mRNA was reverse transcribed to first strand cDNA using Ion Total RNA-Seq kit v2 (Life Technologies, USA). All the four samples (control and induced of each fungus) were systematically labelled with different Ion Xpress™ RNA 3′ barcode primers to make barcoded libraries and second strand cDNA was synthesized using Platinum PCR SuperMix, a high fidelity mix of Ion Total RNA-Seq kit v2 (Life Technologies, USA). Size distribution of the library was evaluated by High Sensitivity DNA Analysis kit on Agilent 2100 bioanalyzer. Emulsion based clonal amplification and enrichment of the template-positive Ion Sphere Particles (ISPs) was done using Ion OneTouch™ 200 template kit v2 DL on Ion OneTouch™ two system and ion one touch ES respectively (Life Technologies, USA). Sequencing was carried out on Ion Personal Genome Machine (PGM™) version 3.0 with Ion PGM™ sequencing 300 bp chemistry and Ion 316™ chip.

De novo assembly of reads and differential gene expression analysis

The raw reads obtained from sequencing run were filtered for minimum length of 40 bases and mean quality score Q ≥ 20 using PRINSEQ 0.20.2 (Schmieder and Edwards 2011). The reads of induced samples were mapped to Caenorhabditis elegans whole genome and mapped reads were eliminated. To build a reference file for expression analysis, all the good quality reads of both the fungi were pooled together to ensure maximum coverage of the probable genes. Good quality reads were assembled into non-contiguous contigs using CLC genomic workbench version 4.9 (CLC Bio, Denmark), where parameters were set as minimum contig length of 200 bp and minimum 80% overlap with 90% identity among the reads. BLASTX search of all the assembled contigs were performed against non-redundant (nr) database at NCBI with minimum E-value 1E-6 using BLAST2GO (Gotz et al. 2008) and results were further utilized to prepare a reference for gene expression analysis using costumed script.

In our previous study (Pandit et al. 2014b), A. conoides was found superior than D. flagrans in catching and killing the root-knot nematodes hence, our main focus of analysis was to make out the virulence genes of A. conoides. Although, we had also studied differentially expressed genes in D. flagrans. Differentially expressed genes were identified using RNA-Seq analysis application provided by CLC genomic workbench V4.9. Expression values were calculated in terms of Reads per Kilobase per Million mapped reads (RPKM). Statistical analysis was carried out in CLC genomic workbench. Expression values (RPKM) were normalized using scaling method (Robinson and Oshlack 2010) and transformed on log2 base scale. Functional annotations of the expressed genes i.e. BLASTX, GO, enzyme code, InterProScan and KEGG pathway analysis was performed using Blast2GO, where E-value-hit-filter was set to 1E-6. GO enrichment analysis was carried out on resulting GO annotations using Fisher’s exact test (p-value < 0.001) to find out the significant categories in induced fungus. Differentially expressed genes were categorised in different groups according to the sequential events that takes place during trapping and digestion of nematodes. The differential expression of genes homologous to pathogen-host interactions (PHI) database was also carried out by mapping the reads with PHI database V 4.0. For comparative orthologous analysis with other fungi, predicted proteins sequences (7225) of A. conoides were compared with protein sequences of other three nematode-trapping fungi with diverse morphology of trapping structures using OrthoVenn (Wang et al. 2015) and Hypergeometric test was used for GO enrichment (p-value < 0.05).

Validation of sequencing data using qRT-PCR

In order to validate the RNA-Seq data, qRT-PCR of 20 selected differentially expressed genes in A. conoides were used. To design the primer for target genes, the assembled contigs of each unigenes was search for blastx and the conserved sequences were subjected to Batch Primer3 v1.0 (You et al. 2008). Primers used for the qRT-PCR study are listed in supplementary table T1. RNA was re-isolated from preserved samples and approximately 1 µg of total RNA was used for cDNA synthesis using Transcriptor First Strand cDNA Synthesis Kit (Roche, USA). qRT-PCR was performed with the help of SYBR Green Real-Time PCR Kit (Life Technologies). 15 µl PCR reaction mixture contains 2 µl cDNA, 0.25 µl each of forward and reverse primer (10 µM), 7.5 µl of 2X SYBR green mix and 5 µl water. A one step thermal cycler programme used for experiments as follow: 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, and 60 °C for 1 min. This was further followed by melt curve to assess the specificity of the primers. Three genes viz. G6PD, G3PDH and glycerol kinase were used as an endogenous control to normalize the data. Three replicates of sample for each primer set were taken and fold change in gene expression was calculated using 2−△△CT method and results were compared with RNA-Seq data.

Results

Both fungi were baited against the nematodes crude extract for 24 h followed by observation under the microscope. A. conoides makes use of an adhesive network of hyphae to clutch the prey nematodes and noticeably demonstrated move from saprophytic to predacious life style after induction. After 24 h, A. conoides successfully developed abundant trapping structures compared with control. However, in case of D. flagrans the induction was poor or not at all, in terms of formation of trapping devices. Figure 1 shows morphological characteristics of trapping devices of A. conoides and D. flagrans. In agreement with our previous study (Pandit et al. 2014b), the results confirm that A. conoides get induced when nematodes are in the vicinity.

Fig. 1
figure 1

Morphological characteristics of trapping structures of A. conoides (a, b and c) and D. flagrans (e and f) on half strength CMA. a, b A characteristics trapping structures of A. conoides under 10× and 40× respectively after induction, c trapped nematodes under 10× objective after 24 h incubation. d A characteristics trapping structures of D. flagrans under 40×, e trapped nematodes under 10× objective after 24 h incubation. (Image c and e- are from our separate experiment where we have added live nematodes on growing fungi)

Full size image

Sequencing and de novo assembly of reads

Raw sequences obtained were submitted to NCBI sequence read archive (SRA), BioProject PRJNA238614 and PRJNA239149. Total 2.47 million reads corresponding to 415.3 million bases were retained after removing low quality reads. The filtered reads were de novo assembled to 11,602 non-contagious contigs with N50 444 bp and an average contig length of 422 bp. From these 11,602 contigs, we got hits for 9775 contigs, (9 E-value between 9.7E-06-2 to 6E–180) during BLASTx search against NCBI nr/nt database. An in-house reference (.gbk file) was prepared with 9775 contigs having BLASTx hit using costumed Perl script and subsequently used for RNAseq analysis.

Expression analysis and annotation

Large number of unigenes were differentially expressed (Fold change > ±4) among the control and induced fungi especially in A. conoides (Supplementary Fig. S1). While looking at fold change > ±2, 70 and 56 unigenes were found unique in A. conoides and D. flagrans, respectively while 77 were shared among both the fungi (Fig. 2d). While looking at PHI homologous genes, 80 and 27 unigenes were differentially expressed with fold change > ±1.5 in A. conoides and D. flagrans, respectively (Fig. 2a, b) with only four common in both (Fig. 2c, Supplementary Table T2). Hence, based on the observation of number of trapping structures, comparative genes expression and PHI homologous genes, our major focus was on gene expression in A. conoides for further analysis. A volcano plot clearly illustrates that large number of unigenes are differentially expressed in A. conoides with fold change > ±1.5 (p < 0.01) (Supplementary Fig. S2). Heat maps for important unigenes are shown in Fig. 3, where several unigenes with diverse functions like signal transduction, morphogenesis, hydrolytic enzymes etc. are up-regulated in A. conoides. Considering fold change > ±2.0, 69 and 88 unigenes were up and down regulated, respectively in A. conoides (Supplementary Table T3). Total 1897 transcripts were uniquely expressed (i.e. having RPKM value 0) either in control or in induced fungus, out of these, 805 (Supplementary Table T4) and 1092 (Supplementary Table T5) were unique in control and induced fungus respectively. 888 unigenes were differentially expressed with fold change > ±1.5 (Log2 transformed) out of these, 809 unigenes showed significant difference (p < 0.001) after performing false discovery rate (FDR) test. Of the 888 unigenes, 415 unigenes were up regulated with fold change > 1.5 (Log2 transformed) and 473 are found to be down regulated with fold change >−1.5 (Log2 transformed) in induced fungus.

Fig. 2
figure 2

Heatmap and Venn diagrams for PHI homologous genes (a, b and c) with fold change ≥± 1.5 and d venn diagram for gens fold change ≥ ±2 in both fungi

Full size image
Fig. 3
figure 3

Heatmap of some important differentially expressed unigenes in A. conoides after induction. a Morphogesis, cell cycle control and pathogenesis related genes, b hydrolytic enzymes and c signal transduction and stress response related genes

Full size image

Further, transcripts from the control and induced samples of A. conoides were de novo assembled to non-contiguous contigs and annotated using Blast2GO. Best BLAST hits were obtained with E-value ranging from 1E-10 to 1E-50. Top hit species distribution graph showed the best hits with Arthrobotrys oligospora (a model nematode-trapping fungus), Dactylella haptotyla and Drechslerella stenobrocha (Fig. 4). GO annotation of the sequences were assigned to three basic categories, biological process (BP), molecular function (MF) and cellular component (CC). Using Fisher’s exact test, enriched GO were extracted. Within category BP, tricarboxylic acid metabolic process (GO:0072350), tricarboxylic acid cycle (GO:0006099), cellular respiration (GO:0045333), citrate metabolic process (GO:0006101) and aerobic respiration (GO:0009060) were enriched in induced fungus. Among MF, genes for the nucleic acid binding (GO:0003676), nucleotide binding (GO:0000166), organic cyclic compound binding (GO:0097159), nucleoside phosphate binding (GO:1901265) and heterocyclic compound binding (GO:1901363) were abundant in induced fungus compare to un-induced one (Fig. 5, Supplementary Table T6). While in CC, no significant category was enriched in induced fungus with p-value < 0.001, however, cellular bud (GO:0005933), respiratory chain complex (GO:0098803), 1,3-beta-D-glucan synthase complex (GO:0000148), nuclear chromatin (GO:0000790) etc. were enriched with p-value < 0.05. These genes might be involved in the development of trapping structures and other morphological changes. On the other hand, in control fungus, only cytoplasmic part (GO:0044444), cytosol (GO:0005829), cytoplasm (GO:0005737), carbon–carbon lyase activity (GO:0016830) and cellular metabolic compound salvage (GO:0043094) categories were enriched with p-value < 0.001 (Fig. 5, Supplementary Table T6). Enzyme abundance was also different, oxidoreductaes and hydrolases group of enzymes were abundant in induced fungus while there was paucity of lyases. No variations were observed with isomerases and ligases class (Supplementary Fig S3). Among the top 15 KEGG pathways, enzymes associated with the TCA cycle, glycolysis, oxidative phosphorylation, methane, nitrogen, pyruvate, sucrose and starch metabolism were abundant in induced fungus, while as the enzymes associated with purine and pyrimidine metabolism were abundant in control fungus (Supplementary Fig S4).

Fig. 4
figure 4

Top-hit species graph from BLASTX. BLASTx was performed against NCBI nr database with minimum E-value 1E-6 using Blast2GO

Full size image
Fig. 5
figure 5

Enriched gene ontology in induced and control using Fisher’s exact test with p-value < 0.001. The graph was generated using Blast2GO

Full size image

We found 1152 orthologous gene clusters among all the four nematode-trapping fungi. 87 clusters were found unique for A. conoides (Fig. 6). Hypermetric enrichment extracted fungus specific GOs where highest, 11 GOs were enriched in A. conoides, majority of which were amino acid transporters and fungal-type cell wall beta-glucan metabolic process (Table 1). While comparing the homologous gene clusters, 101 and 43 were found unique in control and induced library of A. conoides, respectively.

Fig. 6
figure 6

Orthologous gene clusters in A. conoides and other three nematode-trapping fungi

Full size image
Table 1 Hypergeometric GO enrichment analysis of orthologous genes in for nematode-trapping fungi
Full size table

Validation of RNA-seq data using qRT-PCR

Among the three reference (endogenous control) genes, G6PD was found to be steadily expressed in both control and induced fungus, hence it was used for normalization. Of the 20 selected differentially expressed genes, majority of the unigenes showed similar profile for both qRT-PCR and RNAseq data (Fig. 7, Supplementary Table T7). Few genes showed different pattern compared to the RNA-Seq data, which may be due to technology variation (RNA-Seq vs qRT-PCR) and not using the same RNA for RNA-Seq and qRT-PCR.

Fig. 7
figure 7

Comparison of RNA-Seq versus qRT-PCR data. The graph showed the comparative fold change in RNA-Seq and qRT-PCR

Full size image

Various genes involved in the morphogenesis, cell division (Fig. 2a; Table 2), hydrolytic enzymes (Fig. 2b; Table 2), signalling pathways, stress response (Fig. 2c., Table 2), energy production and conservation (Supplementary Table T8), were found up regulated in an induced sample of A. conoides. A proposed model for signalling in A. conoides is shown in Fig. 8 and proposed model of catching and killing of nematodes by A. conoides is demonstrated in Fig. 9. This model is somewhat similar to the previously described model for A. oligospora (Yang et al. 2011b) and general model described by (Li et al. 2015). In addition to all genes described, 30 hypothetical proteins were also found to be differentially expressed with fold change ±2.0 (Supplementary Table T9).

Table 2 Differential expression of some important genes in A. conoides after induction with nematode extract for 24 h
Full size table
Fig. 8
figure 8

A proposed model of signalling in the A. conoides. Expression value of each unigenes in given red in color in bracket. *Indicate fold change in qRT-PCR. (Color figure online)

Full size image
Fig. 9
figure 9

A proposed model of catching and killing of Meloidogyne sp. of root-knot nematodes by A. conoides

Full size image

Discussion

Nematode-trapping fungi are of scientific interest as it can serve as bio-control agents for nematode pest in the agriculture. A. conoides is a nematode-trapping fungi which successfully trap nematodes by means of sticky ring network made up of mycelia (Saxena et al. 1987). It usually survives as a saprophyte in the soil, however, when nematodes are nearby, it transforms into predacious fungus and have a nutritious diet. As we have seen in this study as well noticed in our previous studies, presence of prey makes morphological progressions in A. conoides. Therefore, it is very essential to understand the molecular processes underlying this shift, in order to successfully employ it in a biocontrol programme. RNASeq is widely used method to study gene expression in various forms of life. In the present study, efforts are made to recognize the fundamental genes and their possible role in A. conoides during infection in the nematodes.

Compared to control, abundant trapping structures were produced in A. conoides, after 24 h of incubation along with the nematodes extract. Though we noticed very few trapping structures on the control plates as well, however, numbers were significantly quite less and they were not much developed as observed after induction. Whereas, with D. flagrans no major changes were observed in the mycelia morphology on the plates. This clearly elucidate the host specificity of nematode-trapping fungi and it was in the agreement with (Andersson et al. 2014). The induction in A. conoides may be in response to the pheromones present in the nematodes extract (Hsueh et al. 2013). A smaller amount in fold change (<1.5 fold) with some unigenes could be explained by the induction time (24 h). This was previously observed by (Liang et al. 2013) while studying the cell wall related proteins in A. oligospora. Still in A. conoides, most of the significant genes showed up regulation even after 24 h. Many of the up-regulated unigenes viz. MAP kinases, autophagy proteins and hydrolases are similar to pathogenicity associated genes often found in A. oligospora, M. haptotylum, M. lysipagum, P. chlamydosporia and D. stenobrocha (Ahren et al. 2005; Andersson et al. 2013; Khan et al. 2008; Larriba et al. 2014; Liu et al. 2014; Meerupati et al. 2013; Yang et al. 2011b). Moreover, we also proposed a signalling pathway in A. conoides as well a model for the entire mechanism of catching and killing of nematodes. The general model for entire process is somewhat similar to the reports of (Yang et al. 2011b). This study elucidates involvement of diverse genes of nematophagous fungi during the predation and infection to nematodes. Top BLASTX hits are homologous to A. oligospora and fungal plant pathogens. As A. conoides is a predator or in other sense pathogenic to root-knot nematodes such genes are very crucial during the predacious phase.

The de novo assembled 11,602 contigs from both the fungus libraries, which were used as reference for RNA-Seq, is very close to number of predicated genes reported in other nematode-trapping fungi A. oligospora (11,479) (Yang et al. 2011b) and Monacrosporium haptotylum (10,959) (Meerupati et al. 2013). This shows that reference which was used in expression analysis is acceptable to explain all the genes. The differentially expressed GOs in the category of BP and MF described the genes involves in different cellular processes, required for rapid cell growth and energy metabolism. Whereas in case of CC, enriched GOs are associated with morphogenesis, differentiation and development during the induction. This response might be due to the pheromones (ascarocides) (Hsueh et al. 2013) or volatile organic compounds (Li et al. 2015) and thus stimulated the fungus to produce a large number of trapping structures which eventually help to catch the nematodes. Once nematodes are trapped inside the trapping networks, A. conoides secrets extracellular enzymes to invade the host cuticle and simultaneously take up nutrients from the host which is evident from presence of high numbers of hydrolases during the induction. Of the unigenes with >twofold change, majority of them are hypothetical proteins and genes contributing in morphogenesis, signalling and hydrolytic enzymes. Enrichment of catalytic activity is owing to the fact that the fungus has to metabolize proteinous and other molecules accumulated from the host, which is only possible with activities of different enzymes. Similar is the case for the hydrolases and oxidoreductase group of enzymes, which are involved in the degradation of macromolecules and stress response respectively. Expression of PHI homologous genes farther aid up during the entire process. Enrichment of amino acid transport related orthologous genes in A. conoides explains that this fungus rapidly metabolize nematode constituents especially proteins.

According to sequential events which take place in nematode-trapping fungi, the differentially expressed unigenes are characterized into five different categories and their participation at each particular stage is discuss below.

Signaling process in A. conoides

Initially ‘Nemin’, a peptide secreted by nematodes was thought as an external signalling molecule (Pramer and Stoll 1959). Recently, it has been identified that nematode pheromones (Ascarocides) triggers trap formation in A. oligospora (Hsueh et al. 2013). It is widely known that a signal transduction cascade regulates fungal development and its pathogenicity (Lengeler et al. 2000). The results of our study revealed that G-proteins plus MAPKs play a very vital role in A. conoides, specially in sensing the presence of its prey and in other related signaling events. Proposed model for signaling (Fig. 5) where G-protein coupled receptor, G-protein subunits (α, β & γ), mitogen activated protein kinase (MAPK) mfa1, mitogen activated protein kinase kinase (MAPKK), MAP kinase kinase kinase, serine threonine protein kinases srk1, psk1 and chk2 were up-regulated during induction. Moreover, MAPK pathways assimilate the signals from multiple receptor pathways including two-component signalling systems (Gustin et al. 1998). Histidine kinases are the part of basic two component system by which some organisms sense and respond to the environment (Catlett et al. 2003). In A. conoides, histidine kinase hhk2p and histidine kinase m232p were up-regulated in the presence of nematodes extract. This suggests that MAPKs and histidine kinases are vital for signalling in A. conoides. Our results also correlates with the signal transduction pathways in A. oligospora which gets up regulated after 10 h of incubation with nematode extract (Yang et al. 2011b) and pathogenicity related genes in M. haptotylum (Meerupati et al. 2013), however, in both these fungi the role of G-proteins does not exists in signalling. The role of G-protein signaling is described in D. stenobrocha when the nematode enters into constricting ring (Meerupati et al. 2013). Thus, various differentially expressed genes of signal transduction pathways in A. conoides have proposed function in sensing the nematodes and for switching to predacious life form, which additionally add up in stress response later on.

Genes involved in morphogenesis and cytoskeleton building, essential to build up trapping structures

Next crucial step for infection is formation of hyphal trapping devices, which help fungus to trap the prey nematodes. A. conoides undergoes different cellular and morphological alterations during the process of trapping and digesting the prey nematodes. One of the key features is the formation of trapping devices, which require proteins for cell division, cell wall biogenesis, morphogenesis etc. Several key genes for morphogenesis were found to be up regulated in A. conoides in the presence of nematode extract. Enzymes involved in cell wall synthesis viz., chitin syntheses, chitin synthase activator and glucan synthase were up regulated in induced fungus. In addition, cell wall modifying enzymes, endo-β-glucanase, GPI enchored endo- β-glucanase and endoglucanse-II, which might have role in morphogenesis of trapping devices, were also found up regulated in the induced fungus. Further, a GO:0070879 which is involved in fungal-type cell wall beta-glucan metabolic process was unique to A. conoides and its homologs were absent in other nematode-trapping fungi. Astonishingly, a GPI-anchored cell wall β-endo-glucanase which is also a crucial enzyme for cell wall synthesis was down regulated in the induced fungus.

An autophagy related proteins and autophagy regulatory proteins were found up regulated in the presence of nematode extract in A. conoides. Autophagy is required for trap formation in A. oligospora (Chen et al. 2013). It has been widely accepted that nutrients deprived conditions and presence of nematodes triggers nematophagous fungi for predaceous life, this could be the reason for up regulation of autophagy proteins in A. conoides, as we have used half strength medium during the experiments. This was also supported by enrichment of amino acid transports in A. conoides as generally under nutritional deprived conditions nematophagous fungi use nematode as source of nitrogen if they are present in the vicinity. Besides these genes, cofilin cof1, Ras and Rac were also differentially expressed in A. conoides when baited with nematode extract. All these proteins have proven role in the morphogenesis, especially in dimorphic fungi (Ballou et al. 2013; Banuett et al. 2008; Berepiki et al. 2011; Fortwendel 2015), suggest their role in morphogenesis and trap formation in A. conoides as well. Interestingly, a Myb family of conidiophores development protein was also expressed in A. conoides in the presence of nematode extract. Though the data are not presented here, A. conoides produced large number of conidiospores when incubated with live nematodes for 24 h. Thus, up regulation of this gene in A. conoides explains the production of bulky conidiospores. In addition, capsule polysaccharide biosynthesis protein and capsular associated proteins were differentially expressed in the induced fungus. These proteins may have role in formation of outer adhesive layers on the network of trapping device which aid in trapping the nematodes. A capsular associated protein might have similar function that of AOL protein of A. oligospora (Rosen et al. 1992). Septin, a very essential protein associated with cytokinasis, cell polarity, morphogenesis (Alvarez-Tabares and Perez-Martin 2010) was also found up regulated in A. conoides in the presence of nematode extract. In addition, several transcription, translation and cell division control proteins were also unregulated in induced fungus. These genes contribute in fungus growth, development and also in nutrient acquisition. All the differentially expressed genes altogether are required for development of trapping devices.

Peroxisomal proteins which are found in dense bodies of trapping structure

Trap cells of nematode-trapping fungi possess electron dense bodies which are common in functions like peroxisome (Dowsett et al. 1977; Heintz and Pramer 1972; Veenhuis et al. 1989). In A. conoides, several peroxisome genes, peroxisome biosynthesis protein (pas1), peroxisomal dehydratase, peroxisomal catalase, peroxisomal matrix protein, peroxisomal hydratase-dehydrogenase-epimerase and peroxisomal targeting signal-1 receptor were up regulated in the presence of nematode extract. In addition, a woronin body major protein which are involved in formation of woronin bodies and are fungal peroxisome was also found up regulated in A. conoides. Woronin bodies are involved in healing of mycelia and pathogenesis in various plant pathogenic fungi (Beck and Ebel 2013; Soundararajan et al. 2004), as well as also reported for pathogenicity in A. oligospora (Yang et al. 2011a). Thus these genes may have similar role in A. conoides. Peroxisomal catalase may be in involved in oxidative stress response, pathogenicity and other functions as reported in plant and entomopathogenic fungi (Chantasingh et al. 2013; Paris et al. 2003). Similarly, the malate dehydrogenase, isocitrate lyase and malate synthase, enzymes of the glyoxylate cycle were also found up regulated in induced fungus and glyoxylate cycle has been reported for the fungal virulence (Lorenz and Fink 2001) and also for pathogenesis in A. oligospora (Zhao et al. 2014).

Hydrolytic enzymes vital to degrade the host cuticle and other polysaccharides

Once the target nematodes are trapped into the trapping devices, next course of action is to penetrate host cuticle, where cuticle is degraded to acquire the nutrients. Extracellular hydrolytic enzyme plays a decisive role in degradation of exterior cuticle of host, as it is made up of chitin and several proteinous materials (Yang et al. 2007c). Till date, several pathogenesis associated proteases (Braga et al. 2012; Cruz et al. 2015; Nagee et al. 2008; Wang et al. 2009; Yang et al. 2005, 2007a, b, 2011c) and chitinases (Gan et al. 2007; Nguyen et al. 2008; Tikhonov et al. 2002; Yang et al. 2010) have been reported in numerous nematophagous fungi. In A. conoides, subtilicin like serine protease, alkaline serine protease, ATP dependent protease la, ADAM-B protease, small secreted protein (SSPs) and cell wall associated hydrolase were up regulated in the presence of nematode extract. We have already confirmed the induction of serine protease gene in our fungus using real time PCR earlier (Pandit et al. 2014a) and in this study too. Small secreted proteins play an imperative role in virulence of plant pathogenic fungi, similar may be its function in A. conoides. In addition, a class iii chitinase was up regulated in induced A. conoides. Chitinases are involved in the degradation of the egg shell in egg parasitic fungi (Khan et al. 2004; Tikhonov et al. 2002; Van Nguyen et al. 2007). Thus, proteases and chitinases together play a very key role in nematode-trapping fungi, during the process of infection and simultaneous degradation of the nematode cuticle.

Stress responsive and other genes

Owing to their sessile life, fungi have developed sophisticated responses to the diverse environmental stresses (Kroll et al. 2013). As A. conoides invade host, it gets exposed to innumerable host metabolites and other bio-molecules which might function as armoury for the defence of host. Stress responsive genes, thioredoxin reductase, hsp70 chaperone, nitric oxide reductase, general stress response protein, stress response rci and multi-sensor hybrid histidine kinase were found to be up regulated in A. conoides in the presence of nematode extract. Other unigenes described earlier are also found to be involved in the stress responses. In a variety of fungi, the adjustment of transcriptome has been reported during invasion into the host (Cairns et al. 2010). The function of oxidative pathways, thioredoxin reductase and nitric oxide reductase to escape oxidative stress is reported in variety of pathogenic fungi (Brown et al. 2009; Missall and Lodge 2005; Thon et al. 2007). So we may conclude that A. conoides not only kills and acquires the nutrients from the host but also protects themselves from extraneous environmental stresses. In addition, the differentially expressed hypothetical proteins may have played important roles during the predation, however currently no description is available hence, their functions need to be understood further. Although this study provides new insights and help to understand the fungal parasitism, further studies are required to understand the precise mechanism. Further, the genome information is unavailable for A. conoides, the exact gene annotation is bit complex with de novo approach. Hence, in future the genome sequencing and transcriptome analysis at different time interval may help to understand the process in more detail.

Thus, the current data delivers a comprehensive outlook of differentially expressed genes in A. conoides both in the presence and absence of the host nematode (Meloidogyne sp.) extract. The differentially expressed gene seems to be pathogenicity related genes, often found in A. oligospora and plant pathogenic fungi. As A. conoides is a dimorphic fungus, it shows its predacious personality, this may be a core reason of sharing more genes with pathogenic fungi, however, not harming plants anyways. Genes of signal transduction pathways including G-proteins, MAPKs and histidine kinases may be initial players to recognise the presence of its prey. Distinct classes of chitinases, capsular polysaccharide biosynthesis protein and genes associated with morphogenesis, which are essential for the development of trapping structures and ultimately to invade inside the host were also up regulated. Similarly, peroxisomal biosynthesis proteins, catalase, oxidase and major enzymes of the glyoxylate cycle which are critical to combat oxidative stress and also involved in pathogenicity were found to be up regulated. Hydrolytic enzymes including subtilicin like serine proteases, class iii chitinase, metalloprotease and small secreted proteins may be involved in digesting nematode cuticle to dig up nutrition from the host. As fungi degrade host cuticle, it is subjected to the oxidative and osmotic stress due to host metabolites and defence machinery. Besides all thioredoxin reductase, nitric oxide reductase, heat shock proteins, molecular chaperone and some additional proteins were differentially expressed to evade this stress. One can also make out the functions of differentially expressed hypothetical genes according to (Shen et al. 2015). Thus, our study enhances the knowledge about the fungal parasitism and will further help to understand and improve the nematode-trapping efficiency, leading to development of superior biocontrol agent against the root-knot nematodes. Still the combined approach of using genomics and proteomics could further help to understand the mechanism of A. conoides in detail.